协同感知系统中一种基于信息年龄优化的服务策略

A Service Strategy Based on Age of Information Optimization in Collaborative Perception Systems

边缘计算(EC)技术通过在网络边缘实时处理数据,解决了单车感知、传感器处理和传输延迟等问题,为提供高效、安全的自动驾驶服务提供支持。时敏信息是自动驾驶的核心问题,信息年龄(AoI)成为解决实时性和性能问题的关键指标。将AoI引入自动驾驶EC场景,在协同感知系统架构中,以AoI为主要优化目标,提出最大信息年龄优先的服务策略。通过计算服务策略的时间指标理论值,寻找影响系统性能的关键参数。采用蒙特卡洛方法对传统服务策略和所提的策略进行对比实验。仿真结果表明,在随机初始化的批次结构下,所提的服务策略具有最低的平均AoI,相比并行服务策略降低了54.57%,证明其在AoI优化上的显著优势。

Edge Computing(EC)technology addresses issues such as single-vehicle perception,sensor processing,and transmission latency by processing data in real-time at the network edge,thus providing support for efficient and secure autonomous driving services.Time-sensitive information is a core issue in autonomous driving,and Age of Information(AoI)is emerging as a crucial metric for addressing real-time and performance concerns.This study introduces the concept of AoI in the EC scenario of autonomous driving.Within the architecture of collaborative perception systems,this study proposes a service strategy that prioritizes maximum AoI as the primary optimization goal.Subsequently,by calculating the theoretical values of the time indicators for the service strategy,the key parameters affecting system performance are identified.Finally,using Monte Carlo methods,comparative experiments are conducted between conventional service strategies and the method proposed in this study.Simulation results indicate that under randomly initialized batch structures,the proposed service policy exhibits the lowest age of information,reducing it by 54.57%compared with the parallel service policy,demonstrating its significant advantage in AoI optimization.